Image Courtesy: Getty Images
A proposed plan tied to SpaceX’s long-term vision of deploying up to one million AI-powered satellites in orbit is drawing sharp criticism from some of the space industry’s most respected figures, who argue the idea could prove financially unsustainable despite growing enthusiasm around artificial intelligence.
The debate centers on the concept of turning space into a vast network of orbital data centers capable of running advanced AI models. Critics say the scale, cost, and technical hurdles involved would make such a project extraordinarily difficult to justify economically, especially in the near term, according to Forbes.
Among the most vocal skeptics is aerospace engineer Robert Zubrin, founder of the Mars Society and a longtime supporter of Elon Musk’s ambitions in space. Zubrin argues that deploying a million orbital AI satellites could cost trillions of dollars, potentially rivaling the projected value of SpaceX itself.
The criticism comes despite SpaceX’s remarkable success with its Starlink network, which has already grown to roughly 10,000 satellites and transformed the economics of satellite internet. However, analysts note that broadband satellites provide a clear service that requires space-based infrastructure, while AI computing can currently be performed far more cheaply on Earth.
At the heart of the debate is energy. Orbital AI data centers would rely largely on solar power, but experts argue that generating computing power in space remains significantly more expensive than using terrestrial energy sources such as grid-scale solar, natural gas, hydroelectricity, or nuclear power.
Supporters of the concept believe that future advances could eventually change the equation. Researchers at Google have recently explored theoretical designs for large-scale space-based AI infrastructure, suggesting that dramatically lower launch costs and fully reusable rockets could make orbital computing economically competitive sometime in the mid-2030s.
Those researchers emphasize that such systems remain a long-term “moonshot” rather than an immediate commercial opportunity. Their projections assume substantial reductions in launch costs, major improvements in spacecraft manufacturing, and a much higher launch cadence than exists today.
SpaceX’s own roadmap depends heavily on the success of Starship, the company’s next-generation fully reusable rocket. While recent test flights have demonstrated progress, the vehicle is still undergoing development, and experts question whether SpaceX can realistically scale operations from its current launch rate to the thousands of annual flights that a mega-constellation of orbital AI data centers would require.
The discussion highlights a broader trend emerging across the technology sector. As demand for AI computing continues to surge, companies are increasingly exploring unconventional ways to secure energy, processing power, and infrastructure. While orbital data centers may one day become technically viable, many industry observers believe the economics remain years, if not decades, away from supporting deployment at massive scale.
